Incidents

Why Would You Use Incidents APIs?

The Incidents APIs allow you to raise, retrieve, update and resolve data incidents via API. This is useful for raising or resolving data incidents programmatically, for example from Airflow, Prefect, or Dagster DAGs. Incidents are also useful for conditional Circuit Breaking in these pipelines.

Goal Of This Guide

This guide will show you how to raise, retrieve, update and resolve data incidents via API.

Prerequisites

The actor making API calls must have the Edit Incidents privileges for the Tables at hand.

Raise Incident

You can raise a new Data Incident for an existing asset using the following APIs.

mutation raiseIncident {
  raiseIncident(
    input: {
      resourceUrn: "urn:li:dataset:(urn:li:dataPlatform:snowflake,public.prod.purchases,PROD)"
      type: OPERATIONAL
      title: "Data is Delayed"
      description: "Data is delayed on May 15, 2024 because of downtime in the Spark Cluster."
    }
  )
}

Where resourceUrn is the unique identifier for the data asset (dataset, dashboard, chart, data job, or data flow) you want to raise the incident on.

Where supported Incident Types include

OPERATIONAL
FRESHNESS
VOLUME
COLUMN
SQL
DATA_SCHEMA
CUSTOM

If you see the following response, a unique identifier for the new incident will be returned.

{
  "data": {
    "raiseIncident": "urn:li:incident:new-incident-id"
  },
  "extensions": {}
}

Python SDK support coming soon!

Get Incidents For Data Asset

You can use retrieve the incidents and their statuses for a given Data Asset using the following APIs.

query getAssetIncidents {
  dataset(
    urn: "urn:li:dataset:(urn:li:dataPlatform:snowflake,public.prod.purchases,PROD)"
  ) {
    incidents(state: ACTIVE, start: 0, count: 20) {
      start
      count
      total
      incidents {
        urn
        incidentType
        title
        description
        status {
          state
          lastUpdated {
            time
            actor
          }
        }
      }
    }
  }
}

Where you can filter for active incidents by passing the ACTIVE state and resolved incidents by passing the RESOLVED state. This will return all relevant incidents for the dataset.

Python SDK support coming soon!

Resolve Incidents

You can update the status of an incident using the following APIs.

mutation updateIncidentStatus {
  updateIncidentStatus(
    input: {
      state: RESOLVED
      message: "The delayed data issue was resolved at 4:55pm on May 15."
    }
  )
}

You can also reopen an incident by updating the state from RESOLVED to ACTIVE.

If you see the following response, the operation was successful:

{
  "data": {
    "updateIncidentStatus": true
  },
  "extensions": {}
}

Python SDK support coming soon!

3.3 KiB Raw Permalink Blame History